Distributed Time-varying Coordination of Uncertain Nonlinear Multi-agent Systems: A Unified Model Reference Scheme

多智能体系统的分布式协调有许多应用。现有的分布式多智能体协调研究往往局限于静态问题，而时变问题可以响应感兴趣的移动目标和环境/任务的实时变化（如协同探索、协同监测、协同运输、区域跟随编队跟踪和多机器人多目标导航），但更具挑战性。在现实中，多智能体系统经常表现出非线性和不确定性，但相应的结果是有限的。同时解决分布式时变协调和不确定非线性智能体动力学这两个方面的问题，将大大增强多智能体系统的现有功能。不幸的是，由这两个方面引起的相互交织的挑战通常是难以处理的，更不用说由于切换图和部分测量而产生的进一步复杂性。目前还缺乏同时解决这两个问题的系统方法。本课题旨在推导一种新颖的模型参考协调方案，以弥合线性系统的分布式协调与单个不确定非线性系统的跟踪控制之间的差距。教育和外联活动将扩大代表性不足的群体对研究的参与。本课题的目标是结合新的逆回溯设计思想，导出一种新的统一模型参考协调方案，系统地解决不确定非线性多智能体系统的分布式时变协调问题。在该方案中，每个智能体的参考模型被选择为一个线性系统，该系统由邻居之间的物理状态（如相对状态/输出测量）定义的分布式时变非光滑非线性协调项驱动。不确定非线性多智能体系统的分布式时变协调分为三个部分：1)参考模型本身的协调；Ii)采用不确定非线性动力学方法跟踪各智能体参考模型的状态/输出；Iii)前两部分的相互影响/反馈研究。PI将探索该方案的使用，以适应各种不确定性和动态，并通过考虑不可用信息和不同类型和顺序的异构不确定系统的非光滑非线性算法解决分布式时变协调。所提出的研究将解决具有挑战性的开放研究问题，并推动分布式多智能体协调的发展。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Distributed coordination of multi-agent systems has numerous applications. Existing research on distributed multi-agent coordination is often limited to static problems, while time-varying problems can respond to moving objects of interest and real-time changes in the environment/missions (for example, coordinated exploration, cooperative monitoring, cooperative transport, region-following formation tracking, and multi-robot multitarget navigation) but are more challenging. In reality, multi-agent systems often exhibit nonlinearity and uncertainty but corresponding results are limited. Addressing the two aspects (distributed time-varying coordination and uncertain nonlinear agent dynamics) simultaneously will substantially enhance the existing functionality of multi-agent systems. Unfortunately, the interwoven challenges caused by the two aspects are often intractable, not to mention further complexity due to switching graphs and partial measurements. A systematic approach to tackle the two aspects simultaneously is lacking. This project aims at deriving a novel model reference coordination scheme to bridge the gap between distributed coordination of linear systems and tracking control of a single uncertain nonlinear system. Education and outreach activities will broaden participation of underrepresented groups in research.The objective of this project is to derive a novel unified model reference coordination scheme combined with new inverse backstepping thinking in design to systematically address distributed time-varying coordination of uncertain nonlinear multi-agent systems. In the proposed scheme, each agent’s reference model is chosen as a linear system driven by distributed time-varying nonsmooth nonlinear coordination terms defined by physical states between neighbors (e.g., relative state/output measurements). Distributed time-varying coordination of uncertain nonlinear multi-agent systems is divided into three parts: i) coordination of the reference models themselves; ii) tracking of the state/output of each agent's reference model by with uncertain nonlinear dynamics; iii) study of mutual influence/feedback of the first two parts. The PI will explore the usage of the scheme to accommodate various uncertainties and dynamics and address distributed time-varying coordination with nonsmooth nonlinear algorithms accounting for unavailable information and heterogeneous uncertain systems of different types and orders. The proposed research will solve challenging open research problems and advance the state of the art in distributed multi-agent coordination.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.